CNS of remitted animals might be inhibited or altered. signs of EAE around day 16. In experiments designed
This would predict downregulation of cytokine produc- to study CNS T cells from remitted animals, symp-
tion by CNS T cells during remission as has been tomatic mice were selected and allowed to recover
suggested by one report (Merrill et al., 1992). T cells normal function.
retained within the CNS in active disease have been
shown to be of the CD44 high, CD45RB l°w memory/ef- Isolation of LNC and mononuclear cells from CNS
lector phenotype (Jensen et al., 1992; Zeine and Owens, CNS infiltrates were collected by discontinuous den-
1992). CD44, also known as Pgp-1, is a polymorphic sity gradient centrifugation (Zeine and Owens, 1992).
integral membrane glycoprotein (Trowbridge et al., Mice were anaesthetized with chloral hydrate (3.5 g
1982) which has a role in matrix adhesion, lymphocyte kg -~) and perfused through the heart with 100 ml of
activation and lymph node homing, and has been shown PBS. The brains, spinal cords, and LN were then
to be the principal cell surface receptor for hyaluronate collected (brains that were poorly perfused were dis-
(Aruffo et al., 1990). Elevated expression of CD44 is carded), and dissociated by passing through a nylon or
characteristic of memory T cells (Butterfield et al., stainless steel mesh, respectively. The nervous tissue
1989). CD45 is a family of leukocyte specific membrane was centrifuged at 200 × g for 10 min and then resus-
proteins with protein-tyrosine phosphatase activity. pended in 4 ml of 70% isotonic Percoll (Pharmacia,
CD45 isoforms of various molecular masses are pro- Montreal, Quebec) in RPMI 1640 medium. This was
duced by alternative splicing and usage of three exons then overlaid by equal volumes of 37% and 30% iso-
that encode the N-terminal portion of the external tonic Percoll, and the gradient was centrifuged at 500
domain (Barclay et al., 1987; Johnson et al., 1989). × g for 15 min. Mononuclear cells were collected from
Prolonged activation of CD4 + T cells in vitro leads to a the 37% : 70% interface, washed in medium containing
reduction in the level of high M r CD45R expression 10% FCS (ICN Biomedicals) and counted.
(Birkeland et al., 1989). The memory/effector pheno-
type is associated with active cytokine production by T Flow cytometry
cells (Bottomly, 1988). Whether this phenotype is Surface staining for CD4, CD8, CD3,, CD45,
maintained following remission is therefore a question CD45RB, CD44, and TcRa/3 was performed as previ-
that is relevant to the mechanism of remission. ously described (Zeine and Owens; 1992). CD2 expres-
In this study we have isolated CNS mononuclear sion was detected using 12.15A (Altevogt et al., 1989),
cells from mice at various intervals following disease and anti-TCRy/~ mAb was obtained from PharMin-
onset, and used flow cytometry to describe the kinetics gen (San Diego, CA). Where indicated mAbs were
of CD4 + T cell changes in the CNS following remis- purified by Protein G-sepharose affinity chromatogra-
sion, and to determine their surface phenotype. We phy (Pharmacia) and either coupled with biotin by
show loss of CD4 + T cells from the CNS during incubation with biotinamidocaproate N-hydroxysuc-
clinical remission. However, those CD4 + T cells that cinimide ester (Sigma) or fluorescinated by incubation
remained within the neural tissue, maintained an acti- with FITC-Celite (Sigma). Cells (5 × 105-106) were
vated, memory/effector surface phenotype up to 12 incubated with antibody at 4°C for 20 min and then
weeks after the initial attack. washed. Primary rat mAbs that were used as hy-
bridoma supernatants were visualized by using FITC-
goat anti-rat Ig (Southern Biotechnology, Birmingham,
Materials and Methods AL). Biotinylated primary Abs were visualized with
either FITC-coupled streptavidin (Bio-Can Scientific)
Mice or phycoerythrin-coupled streptavidin. Non-specific
Female SJL/J mice (5-8 weeks) were obtained from binding to goat anti-rat Ig was blocked by pre-incuba-
Harlan-Sprague Dawley (Indianapolis, IN). tion with rat Ig (100 /xg ml -~) (Bio-Can Scientific,
Toronto, Ontario), before incubation with PE-CD4
EAE induction, assessment and remission a n d / o r PE-CD8. Surface staining was analysed using a
EAE was induced by s.c. injections on day 0 and day FACScan (Becton Dickinson). Dead cells were ex-
7, of either 0.5 mg rat spinal cord homogenate (RSCH) cluded by propidium iodide staining.
or 400 /xg bovine myelin basic protein (MBP) (Sigma,
St. Louis, MO) in CFA (50 ~g Mycobacterium tubercu-
losis H37RA (Difco, Detroit, MI) per mouse). Mice Results
were monitored daily for symptoms and assigned clini-
cal scores as follows: 0 (no signs), 1 (flaccid tail, clumsi- Correlation between clinical remission and the number of
ness), 2 (moderate paresis), 3 (severe paresis or unilat- CNS CD4 + cells
eral hind limb paralysis), 4 (bilateral hindlimb paraly- Mononuclear cells were isolated by discontinuous
sis). Between 50 and 70% of animals developed clinical density gradient centrifugation (Zeine and Owens,
1992) from the pooled brains and spinal cords of SJL//J
female mice. During the active phase of E A E (day 17
post-immunization), the proportion of the CNS 10 3
mononuclear cells that were CD3 ÷ and expressed high CD8
levels of CD4 was 4-5-fold greater than in remitted or 10 2
naive mice (Fig. 1). The small proportion of CD3 ÷
C D 4 - cells seen in Fig. 1, represents CD8 ÷ T ceils,
which did not increase in remitted mice (Fig. 1). Maxi-
mal CD8 to CD4 ratios within the CNS were obtained
during the active phase of E A E and did not exceed 101 10 2 10 3
0.37 (Fig. 2). This and previous work show that CD4 ÷
and CD8 ÷ ceils isolated from CNS are all CD3 ÷ T C D 3
cells. Macrophages/microglia could be excluded from Fig. 2. Isolation of CD8 + T cells from the CNS of mice with active
EAE. CNS mononuclear cells were isolated from ten mice in the
analysis by their 10-fold lower expression of CD4 active phase of E A E (day 17 post-immunization), and stained with
(Sedgwick et al., 1991). biotinylated-anti-CD8 and FITC-CD3. Anti-CD8 was visualized with
phycoerythrin-coupled streptavidin. The figure shows CD8 plotted
against CD3 expression on CNS mononuclear cells.
In order to assess the state of CNS infiltration
CD4 I "'?-" "~" "
EAE during clinical remission, groups of mice with E A E
were allowed to recover normal motor function and the
10 2 number and phenotype of CNS-derived mononuclear
cells were analyzed through remission. The number of
101 cells that could be obtained from individual mice was
less than required for flow cytometric analysis, so
pooled samples from groups of ten mice were analysed.
A strong correlation between number of CD4 ÷ T cells
and disease progression was observed. The mean num-
I ber of mononuclear cells obtained per mouse CNS was
10 3 I
I NAIVE only slightly reduced in remitted mice (8.6 x 104 -t- 0.2,
' • I ,." :,,:~'.]._ "
n = 4) as compared to symptomatic mice (13.7 x 10 4 -I-
10 2 .__ _ _ _ ~.=S - : ~ . . . . . . . 0.2, n = 3) The absolute number of CD4 ÷ T cells was
.....,+.: :+++24-,+ " .
dramatically reduced from 77.1 x 103 on day 17 to
101 2.6 x 103 on day 28 (Fig. 3). The proportion of blasts
- :';~N ~.~ • •
~'~ ;i" ~ -': • amongst the CNS CD4 ÷ cells also decreased from 30%
on day 17 to 3% at day 55. The kinetics of CNS
infiltration correlated with the mean clinical scores
i z s%
from groups of ten mice at each time point and in the
experiment done 105 days after immunization there
10 3 i
I • .
REMITTED was a slight rise in both the mean score and the
number of CNS CD4 ÷ cells (Fig. 3). In a previously
1 0 ~. _ :_ _ :., -.a_.a~::-'... "
published study of passively transferred E A E we
: .'~!~ ~ ' " " Z~'~ . " . .
showed that the percent of CD4 + T cells increased
:!:~:: ~;. from 8% 2 days after onset of EAE to 25.8% 3 days
after onset and decreased to 3.5% 6 days after onset
(see Table 1 in Zeine and Owens, 1992). Given that the
101 10 2 10 3 day of onset in active E A E is about 14 days, it can be
appreciated that the results from both active and pas-
Fig. 1. Isolation of CD4 + CD3 ÷ cells from the CNS of naive mice, sive EAE are consistent. It was further possible to
mice with active EAE, and mice in remission. E A E was induced by select cells that expressed a high level of CD4 (T cells)
immunization with R S C H in CFA, and mononuclear cells were for analysis by gating and double staining.
isolated from the CNS of ten mice either at the onset of clinical signs
(day 17 post-immunization) (top) or following remission (bottom).
CNS mononuclear cells were also isolated from naive mice (center).
T cell phenotype of CNS CD4 + cells isolated from
Cells were stained with PE-anti-CD4 and FITC-anti-CD3. The figure clinically remitted mice
shows CD4 plotted against CD3 expression on CNS mononuclear The CD4 ÷ cells isolated from the CNS of mice with
cells. E A E were all CD2 ÷, CD45 +, and CD3 ÷ TCRa/[3 +.
(1.~ et al., 1989). In one study immunocytochemical staining
of CNS frozen sections showed decreased numbers of
(x 10" ~) infiltrating CD4 ÷ T cells in remitted mice (Cannella et
CD4 T 56 al., 1990). The kinetics of T cell loss which we have
cells per now described (Fig. 3) suggest that the majority of
CD4 ÷ T cells that accumulate within the CNS at the
) onset of disease are lost from the CNS within 48 h. A
number of groups have shown that T cells, which are
capable of recognizing a CNS antigen, are either re-
14 l Io,,
Io, (oZ Io)
0 17 18 1; 20 2; 2; 55 1(;5
Days after first immunization I A
Fig. 3. Kinetics of CNS infiltration by CD4 ÷ T cells during EAE. I
E A E was induced in groups of mice by s.c. injections of either R S C H 10 3
or MBP in C F A on days 0 and 7. The mice exhibited signs of clinical
. . . . i -: " "2 ....
E A E between days 16 and 19. After day 20, all the mice had
remitted. In each experiment the brains and spinal cords from ten 10 2
mice were pooled. CNS mononuclear cells were isolated at various - - - ~_~.-,~--=~'-~-7. - - - .
time points and stained with PE-anti-CD4 for analysis by FACS. The
graph represents the n u m b e r of CD4 + T cells obtained per mouse 101
CNS in each experiment. N u m b e r s in parentheses are the mean
clinical scores on the day each experiment was done.
:1 itCg 3%
TCR3~/8 T cells were not detected in the CNS of mice
in the active phase of the disease nor on day 28
post-immunization (not shown).
"::~:i ri" :" :'i .:"
Memory / effector phenotype of CNS CD4 ÷ T cells 101
More than 70% of CNS CD4 ÷ cells, both in active
E A E and in remission, expressed high levels of Pgp-
t ..... 'i~r~g ~, , . . L ...........
1/CD44, whereas less than 20% of LN CD4 ÷ T ceils 200 400 600 800
were CD44 high (Fig. 4A). The level of expression of
CD44 on blasts (defined by forward scatter) was also
high, and was similar between CNS and LN (Fig. 4A). B
The majority of CD4 ÷ cells from LN and blood
expressed high levels of CD45RB. By contrast, more
than 60% of CD4 ÷ T cells in CNS were CD45RB ]°w. J~
The high proportion of CD45RB l°w at disease peak E
was also seen in passively transferred E A E (Zeine and
Owens, 1992) and in other studies (Jensen et al., 1992). =.-
Similar proportions of CD45RB ~°w CD4 ÷ T cells were ¢D
found in naive mice (not shown), mice in the active .>
phase of E A E and in remitted mice (Fig. 4B). Despite 4}
some variability between times of analysis, the propor- |
i ' ' " " 1 ' ' ''""1 '
tion of CNS CD4 ÷ T cells that were CD45RB ~°w re- 10 1 10 2 10 3
mained greater than 60%, and there was no trend in
the variation that could be correlated with disease CD45RB
Fig. 4. P g p - 1 / C D 4 4 and CD45RB expression on CD4 + T cells
progression (Fig. 5).
isolated from CNS and LN of mice in clinical remission. Cells were
isolated from CNS and L N after remission (Day 28 post immuniza-
tion) and stained with either biotinylated anti-CD44 or biotinylated
Discussion anti-CD45RB, the binding of which was visualized using FITC-
streptavidin. PE-anti-CD4 was used to gate on CD4 + cells. (A)
Panels show CD44 plotted against forward scatter for CD4 ÷ cells
Correlation of the onset of clinical signs of E A E from CNS (top) and LN (bottom); (B) Profiles show the distribution
with CNS infiltration by autoreactive helper T cells has of CD45RB expression on CD4 + T cells from CNS (solid line) and
been well documented (Mokhtarian et el., 1984; Lyman LN (stippled line).
I oo tiple sclerosis (Brennan et aI., 1989; Kjeldsen-Kragh et
al., 1990; Viney et al., 1990; Wucherpfennig et al.,
~ 80 1991). y-~ T cells are capable of recognizing antigens
expressed by oligodendrocytes and have been shown to
cause lysis of oligodendrocytes in culture (Freedman et
al., 1991; Selmaj et al., 1992). However, our flow-cyto-
metric analysis of T cells from mice with EAE revealed
no significant proportions of TCRTt~-bearing T cells
within the CNS. 3,-t~ T cells may play a role in chronic
o inflammation, but such conditions are distinct from the
17 23 28 55 105
early stages of EAE that we have studied.
B a y s after immunization
The CD45RB l°w phenotype defines T cells that have
Fig. 5. Proportion of CD45RB l°w CD4 + T cells isolated from the
been activated through antigen recognition (Bottomly,
CNS of EAE and remitted mice. Cells were isolated from CNS of
groups of mice at various times following immunization for EAE
1988). Reversion from CD45RB l°w to CD45RB high has
induction. The cells were double stained with PE-CD4 and anti- been shown to occur (Bell and Sparshot, 1990). One
CD45RB. 23G2 was visualized with either FITC-goat-anti-rat Ig or might predict that T cells in remitted animals would
FITC-streptavidin. Levels of CD45RB expression were defined by not express the CD45RB l°w phenotype, as a conse-
correspondence to the two populations in Fig. 3B. Each histogram
quence either of downregulation by regulatory cells,
shows a separate experiment.
a n d / o r of decreased T C R / C D 3 signalling. Remission
also might be induced by or coincide with the entry to
the CNS of regulatory cells with a naive CD45RB high
tained in the tissue or cyclically re-enter to initiate phenotype. Our results, however, demonstrate that the
a n d / o r perpetuate inflammation (Hickey et al., 1991; majority of CNS CD4 + T cells from remitted mice
Zeine and Owens., 1992). Cells of irrelevant specificity were CD45RB l°w. Indeed, four out of the five groups
were not found within 1-2 days of their entry into the represented in Fig. 4 were in remission and all con-
CNS (Hickey et al., 1991). This argues for antigen tained as high or higher proportions of CD45RB =°w
recognition as a stimulus for T cell retention in the CD4 ÷ T cells as seen at peak EAE. This does not
CNS, but does not explain loss of CD4 ÷ T cells given exclude phenotypic interconversion or immigration of
that myelin antigen concentration does not diminish. It naive cells, but since CD45RB high cells never consti-
has been proposed that suppressor or immunoregula- tuted more than 40% of the CNS T cell populations,
tory CD8 ÷ T cells, a n d / o r the secretion of inhibitory the dynamic equilibrium always favours the activated
cytokines such as TGF-/3 play a role in remission phenotype. The most important difference, however,
(Miller et al., 1991; Jiang et al., 1992; Koh et al., 1992). between mice in the active phase of EAE and remitted
Our results, however, present evidence against any mice was in the number, not in the activation state, of
increase in the number of CNS CD8 ÷ T cells during CNS T cells.
remission phases of EAE (Fig. 1). In summary, we have isolated CD2 ÷ CD45 ÷ CD3 ÷
A role for CD4 ÷ suppressor T cells in the regulation TCRa/3 + CD4 ÷ cells from the CNS of S J L / J mice
of EAE was first demonstrated by the use of suppres- during the remission phase of EAE. There were no
sor cell lines generated in vitro from recovered rats TCRT~ + cells in the CNS, and there was no increase
(Ellerman et al., 1988). Protection against EAE can be in the proportion of CD8 + T cells during remission.
passively transferred by a combination of MBP-primed We have shown a reduction in the numbers of CD4 ÷ T
B cells and a nylon wool adherent subpopulation of cells during remission, but no change in their CD44 high,
CD4 ÷ T cells isolated from recovered rats (Karpus and CD45RB =°w, memory/effector phenotype. These find-
Swanborg, 1991). CD4 + T suppressor cells isolated ings argue against downregulation of T cell function as
from recovered rats had been shown to selectively a mechanism for remission, and instead suggest overt T
inhibit the in vitro production of IFN3, by effector cells cell loss to be the cause.
from rats with EAE (Karpus and Swanborg, 1988). The
CD4 + T cells which we have shown within the CNS of
recovered mice (Fig. 3) could have included suppres- Acknowledgements
sors. However, even if this were the case, they could
not be distinguished from presumed effectors by their We thank Dr. Jia-You Lin for technical assistance
CD45R phenotype. and Dr. Philippe Poussier, at the McGill Nutrition
Elevated proportions of CD3 ÷ TcR78 have been Center in Montreal, for Provision of PE-CD8. This
observed in a number of studies at sites of inflamma- work was funded by The Multiple Sclerosis Society of
tion in some autoimmune diseases such as rheumatoid Canada. T.O. is an MRC Canada Scholar. R.Z. was
arthritis, Sjogren's syndrome, coeliac disease and mul- supported by The Multiple Sclerosis Society of Canada.
References mental autoimmune encephalomyelitis requires both CD4 ÷ T
suppressor cells and myelin basic protein-primed B cells. J.
Altevogt, P., Kohl, U., Von Hoegen, P., Lang, E. and Schirrnacher, Neuroimmunol. 33, 173-177.
V. (1989) Antibody 12-15 cross-reacts with mouse Fc3, receptors Kjeldsen-Kragh, J., Quayle, A. Kalvenes, C., Forre, O., Sorskaar, D.,
and CD2: study of thymus expression, genetic polymorphism and Vinje, O., Thoen, J. and Natvig, J.B. (1990) T3,~ cells in juvenile
biosynthesis of the CD2 protein. Eur. J. Immunol. 19, 341-346. rheumatoid arthritis and rheumatoid arthritis. Scand. J. Im-
Aruffo, A., Stamenkovic, I., Melnick, M., Underhill, C.B. and Seed, munol. 32, 651-660.
B. (1990) CD44 is the principal cell surface receptor for Koh, D.-R., Fung-Leung, W.-P., Ho, A., Gray, D., Acha-Orbea, H.
hyaluronate. Cell 61, 1303-1313. and Mak, T.-W. (1992) Less mortality but more relapses in
Barclay, A.N., Jackson, D.I., Willis, A.C. and Williams, A.F. (1987) Experimental Allergic Encephalomyelitis in CD8 - / - mice. Sci-
Lymphocyte specific heterogeneity in the rat leucocyte common ence 256, 1210-1213.
antigen (T200) is due to differences in polypeptide sequences Lyman, W.D., Abrahams, G.A. and Raine, C.S. (1989) Experimental
near the NH2-terminus. EMBO J. 6, 1259-1264. autoimmune encephalomyelitis: isolation and characterization of
Bell, E.B. and Sparshot, S.M. (1990) Interconversion of CD45R inflammatory cells from the central nervous system. J. Neuroim-
subsets of CD4 T cells in vivo. Nature 348, 163-165. munol. 25, 195-201.
Birkeland, M.L., Johnson, P., Trowbridge, I.S. and PurE, E. (1989) Merrill, J.E., Kono, D.H., Clayton, J., Ando, D.G., Hinton, D.R. and
Changes in CD45 isoform expression accompany antigen-induced Hofman, F.M. (1992) Inflammatory leucocytes and cytokines in
murine T-cell activation. Proc. Natl. Acad. Sci. USA 86, 6734- the peptide-induced disease of experimental allergic en-
6738. cephalomyelitis in SJL and B10.PL mice. Proc. Natl. Acad. Sci.
Bottomly, K. (1988) A functional dichotomy in CD4+ T lympho- USA 89, 574-578.
cytes. Immunol. Today 9, 268-270. Miller, A., Lider, O. and Weiner, H.L. (1991) Antigen-driven by-
Brennan, F., Plater-Zyberk, C., Maini, R.N. and Feldman, M. (1989) stander suppression after oral administration of antigens. J. Exp.
Coordinate expansion of 'fetal type' lymphocytes (TCR gamma Med. 174, 791-798.
d e l t a + T and CD5 ÷ B) in rheumatoid arthritis and primary Mokhtarian, F., McFarlin, D.E. and Raine, C.S. (1984) Adoptive
Sjogren's syndrome. Clin. Exp. Immunol. 77, 175-178. transfer of myelin basic protein-sensitized T cells produces chronic
Butterfield, K., Fathman, C.G. and Budd, R.C. (1989) A subset of relapsing demyelinating disease in mice. Nature 309, 356-358.
memory CD4 + helper T lymphocytes identified by expression of Oppenheim, J.J., Rosenstreich, D.L. and Potter, M. (1981) Cellular
Pgp-1. J. Exp. Med. 169, 1461-1466. functions in immunity and inflammation. Elsevier, New York,
Cannella, B., Cross, A.H. and Raine, C.S. (1990) Upregulation and NY, p. 19.
coexpression of adhesion molecules correlate with relapsing au- Raine, C.S. (1985) Experimental allergic encephalomyelitis. In: J.C.
toimmune demyelination in the central nervous system. J. Exp. Koetsier (Ed.), Handbook of Clinical Neurology. Elsevier, Am-
Med. 172, 1521-1524. sterdam, Vol. 3(47), pp. 429-466.
Ellerman, K.E., Powers, J.M. and Brostoff, S.W. (1988) A suppressor Sedgwick, J.D., Scwender, S., Imrich, H., Dorries, R., Butcher, G.W.,
T-lymphocyte cell line for autoimmune encephalomyelitis. Nature Ter Meulen, V. (1991) Isolation and direct characterization of
331, 265-267. resident microglial cells from the normal and inflamed central
Freedman, M.S., Ruijs, T.C.G., Selin, L.K. and Antel, J.P. (1991) nervous system. Proc. Natl. Acad. Sci. USA 88, 7438-7442.
Peripheral blood 3'-~ T cells lyse fresh human brain-derived Selmaj, K., Brosnan, C.F. and Raine, C.S. (1992) Expression of heat
oligodendrocytes. Ann. Neurol. 30, 794-800. shock protein-65 by oligodendrocytes in vivo and in vitro: Impli-
Hickey, W.F., Hsu, B.L. and Kimura, H. (1991) T-lymphocyte entry cations for multiple sclerosis. Neurology 42, 795-800.
into the central nervous system. Neurosci. Res. 28, 254-260. Trowbridge, I.S., Lesley, J., Schulte, R., Hyman, R. and Trotter, J.
Jensen, M.A., Arnason, B.G.W., Toscas, A. and Noronha, A. (1992) (1982) Biochemical characterization and cellular distribution of a
Preferential increase of IL-2R + CD4 ÷ T cells and CD45RB- polymorphic murine cell-surface glycoprotein expressed on lym-
CD4 ÷ T cells in the central nervous system in experimental phoid tissues. Immunogenetics 15, 299-312.
allergic encephalomyelitis. J. Neuroimmunol. 38, 255-262. Viney, J., MacDonald, T.T. and Spencer, J. (1990) Gamma/delta T
Jiang, H., Zhang, S. and Pernis, B. (1992) Role of CD8 ÷ T cells in cells in the gut epithelium. Gut 31, 841-844.
murine Experimental Allergic Encephalomyelitis. Science 256, Waldor, M.K., Sriram, S., Hardy, R., Herzenberg, L.A., Lanier, L.,
1213-1215. Lim, M. and Steinman, L. (1985) Reversal of experimental aller-
Johnson, P., Greenbaum, L., Bottomly, K. and Trowbridge, I.S. gic encephalomyelitis with monoclonal antibody to a T cell subset
(1989) Identification of the alternatively spliced exons of murine marker. Science 227, 415.
CD45 (T200) required for reactivity with B220 and other T200-re- Wucherpfennig, K.W., Newcombe, J., Cusner, L., Li, H., Keddy, C.,
stricted antibodies. J. Exp. Med. 169, 1179-1184. Weiner, H.L. and Hafler, D.A. (1991) Analysis of aft and y~
Karpus, W.J. and Swanborg, R.H. (1989) CD4 ÷ suppressor cells T-cell receptors in MS plaques. Neurology 41 (Suppl. 1), 380.
differentially affect the production of IFN-~/ by effector cells of Zeine, R. and Owens, T. (1992) Direct demonstration of the infiltra-
experimental autoimmune encephalomyelitis. J. Immunol. 143, tion of murine CNS by Pgp-1/CD44 high CD45RB I°w CD4 + T
3492-3497. cells that induce experimental allergic encephalomyelitis. J. Neu-
Karpus, W.J. and Swanborg, R.H. (1991) Protection against experi- roimmunol. 40, 57-70.